Research And Design On Tire Mark Automatic Recognition System

Automobile tire is an important component to the vehicle. In tire production, once any tire mark happens in error, not only the tire producer will pay huge compensation, tire users may also have accidents in the use. However, currently in the domestic tire manufacturers, it will take a long time to identify the mark of a tire by artificial. If computer assisted identification can be used, the labor time and labor intensity will be reduced greatly, and human and material resources will be saved, there will be a great economic and social benefits.In this paper, Sector Scan method is presented on reading and processing the tire mark image, through researching the structure, queueing shape and the position on the tire of the tire mark. Meanwhile, this method is also used in pretreatment of the tire mark image, such as sectional histogram equalization, gray correction, binarization and so on. Pixel Group Optimize is also put forward to optimize the binarizated image.Secondly, the Tire Mark Domain Positional Algorithm is presented in this paper. This algorithm is based on the Sector Scan. Its positional effection is much more idealized and the domain's edges locking of the tire mark is much more accurate.Thirdly, the algorithm of the second pretreatment to tire mark area increases the specification degree between the text and background of the tire mark.Last, Radial Direction Projection and Arc Direction Projection are used to make projection to the tire mark text. Then draw the projection image. Meanwhile, the algorithm of renewing the text projection's feature according to the optimized projection image is put forward. Afterwards the features of characters are picked up and made the identification. And the result of the identification is good.